Bearing intelligent diagnosis method and system, readable storage medium and computer

Abstract

The invention provides an intelligent bearing diagnosis method and system, a readable storage medium and a computer. The method comprises the steps that a vibration signal, a temperature signal and a sound wave signal of a target equipment bearing are converted into digital signals; inserting a new data point at the time point of each digital signal, and performing signal alignment and filtering noise reduction processing on the processed digital signal to obtain a filtering signal; performing matrix generation on the feature data of each filtering signal to obtain data matrixes, and splicing the data matrixes to obtain a multi-source information fusion tensor; performing dimensionality reduction and feature fusion on the multi-source information fusion tensor by using a tensor decomposition algorithm, and inputting the multi-source information fusion tensor into the long-short-term memory network model for training to obtain a fault diagnosis model; and performing fault diagnosis on the to-be-detected bearing data according to the fault diagnosis model to output a corresponding fault diagnosis result. According to the invention, the obtained multi-source information fusion tensor and the long-short term memory network model are used to construct the fault diagnosis model, and the accuracy and robustness of the fault diagnosis model are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a bearing intelligent diagnosis method, system, readable storage medium and computer. Background Art

[0002] Bearings are one of the most important components in mechanical equipment. They are in a rotating or load-bearing state for a long time. Their operating state has a direct impact on the performance of rotating machinery. Due to their harsh operating environment and the increasing complexity of mechanical structures, they are often prone to failures, causing serious economic losses and even casualties. Common bearing failures include rolling element damage, inner and outer ring cracks, insufficient lubrication, etc. If these failures are not discovered in time, they may cause equipment damage or shutdown.

[0003] At present, most fault diagnosis technologies are relatively simple, often with problems such as low detection accuracy and slow response speed. In addition, the existing monitoring and diagnosis systems cannot be well applied and promoted due to difficulties such as poor stability and difficult maintenance. At present, the most commonly used method for rolling bearing fault diagnosis is to analyze the rolling bearing vibration signal. The data is too simple and often cannot diagnose the fault quickly and accurately. Summary of the invention

[0004] Based on this, the purpose of the present invention is to provide a bearing intelligent diagnosis method, system, readable storage medium and computer to at least solve the deficiencies in the above-mentioned technology.

[0005] The present invention proposes a bearing intelligent diagnosis method, comprising: Collecting vibration signals, temperature signals and acoustic wave signals of the target equipment bearing, and converting the vibration signals, the temperature signals and the acoustic wave signals into corresponding digital signals; Inserting new data points at the time points of each of the digital signals to fill in the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals; Extracting features from each of the filter signals, generating matrices of the extracted feature data, and splicing the obtained data matrices to obtain a multi-source information fusion tensor; Using a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, the multi-source information fusion tensor is input into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; Fault diagnosis is performed on the bearing data to be detected according to the fault diagnosis model to output corresponding fault diagnosis results.

[0006] Furthermore, the steps of collecting the vibration signal, temperature signal and acoustic wave signal of the bearing of the target device and converting the vibration signal, the temperature signal and the acoustic wave signal into corresponding digital signals include: The vibration signal, the temperature signal and the sound wave signal are subjected to signal gain adjustment to obtain corresponding gain signals, wherein the signal gain The corresponding peripheral resistor The relationship is: ; Each of the gain signals is filtered by a filter to obtain a corresponding analog signal, and each of the analog signals is input into a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

[0007] Furthermore, the steps of inserting new data points at the time points of each of the digital signals to fill the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals include: Inserting new data points at the time points of each of the digital signals to fill in the missing data of each of the digital signals; A linear interpolation algorithm is used to connect known data points in the processed digital signal, and estimated values ​​are generated between each of the known data points so that the processed data signal presents a smooth transition on the time axis. The digital signal processed by the linear interpolation algorithm is filtered and denoised to obtain a corresponding filtered signal.

[0008] Furthermore, the calculation formula of the linear interpolation algorithm is: ; In the formula, , represents two known time points in the processed data signal, and ; , Represents the processed data signal at two time points , The value at Indicates the time point where interpolation is required. Represents the signal value calculated by the linear interpolation algorithm.

[0009] Furthermore, the steps of extracting features from each of the filter signals, generating a matrix of the extracted feature data, and splicing the obtained data matrix to obtain a multi-source information fusion tensor include: Using a wavelet packet decomposition algorithm to extract time-frequency domain features of each of the filtered signals, so as to extract the time-frequency features of each of the filtered signals in multiple frequency bands; The time-frequency characteristics of each of the filtered signals in multiple frequency bands are standardized respectively, and the standardized signals are spliced ​​to obtain a multi-source information fusion tensor.

[0010] Furthermore, the calculation formula for the standardization process is: ; ; ; In the formula, Represents the time-frequency characteristics of each filtered signal in multiple frequency bands, Representing time-frequency features The mean of Representing time-frequency features The standard deviation of Representing time-frequency features The number of

[0011] The present invention also provides a bearing intelligent diagnosis system, comprising: A signal acquisition module, used to collect vibration signals, temperature signals and acoustic wave signals of the target equipment bearing, and convert the vibration signals, temperature signals and acoustic wave signals into corresponding digital signals; A signal processing module, used for inserting new data points at the time points of each of the digital signals to fill the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals; A feature extraction module is used to extract features from each of the filter signals, generate a matrix of the extracted feature data, and splice the obtained data matrix to obtain a multi-source information fusion tensor; A model training module is used to use a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, and then input it into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; The fault diagnosis module is used to perform fault diagnosis on the bearing data to be detected according to the fault diagnosis model to output corresponding fault diagnosis results.

[0012] Furthermore, the signal acquisition module includes: A signal gain unit is used to adjust the signal gain of the vibration signal, the temperature signal and the sound wave signal to obtain corresponding gain signals, wherein the signal gain The corresponding peripheral resistor The relationship is: ; The signal conversion unit is used to filter each of the gain signals using a filter to obtain a corresponding analog signal, and input each of the analog signals to a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

[0013] Furthermore, the signal processing module includes: A signal filling unit, used for inserting a new data point at a time point of each of the digital signals to fill in the missing data of each of the digital signals; The signal processing unit is used to connect the known data points in the processed digital signal using a linear interpolation algorithm, and generate estimated values ​​between each of the known data points so that the processed data signal presents a smooth transition on the time axis, and filter and denoise the digital signal processed by the linear interpolation algorithm to obtain a corresponding filtered signal.

[0014] Furthermore, the feature extraction module includes: A feature extraction unit, used to perform time-frequency domain feature extraction on each of the filter signals using a wavelet packet decomposition algorithm, so as to extract the time-frequency features of each of the filter signals in multiple frequency bands; The signal splicing unit is used to standardize the time-frequency characteristics of each of the filtered signals in multiple frequency bands respectively, and to splice the standardized signals to obtain a multi-source information fusion tensor.

[0015] The present invention also provides a readable storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned intelligent bearing diagnosis method is implemented.

[0016] The present invention also proposes a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned intelligent bearing diagnosis method when executing the computer program.

[0017] The bearing intelligent diagnosis method, system, readable storage medium and computer in the present invention use multi-source information fusion tensors to construct a fused feature tensor from multi-source heterogeneous signals such as vibration data signals, temperature data signals and acoustic wave data signals. By fusing fault data with real-time data, the accuracy and sensitivity of fault diagnosis can be effectively improved. The obtained multi-source information fusion tensor is used to construct a fault diagnosis model with a long short-term memory network model to improve the accuracy and robustness of the fault diagnosis model. It can not only reflect the current status of the equipment in a timely manner, but also provide real-time alarms, storage and retrieval analysis, thereby identifying potential faults in advance, reducing downtime, optimizing equipment maintenance cycles, and improving equipment operation efficiency and safety. The diagnosis model is continuously optimized through advanced technologies such as machine learning to adapt to the fault diagnosis needs of different types of equipment, and has wide applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of the bearing intelligent diagnosis method in the first embodiment of the present invention; Figure 2 for Figure 1 Detailed flow chart of step S101; Figure 3 for Figure 1 Detailed flow chart of step S102; Figure 4 for Figure 1 Detailed flow chart of step S103; Figure 5 is a structural block diagram of a bearing intelligent diagnosis system in a second embodiment of the present invention; Figure 6 FIG. 4 is a structural block diagram of a computer in a third embodiment of the present invention.

[0019] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0020] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0022] Embodiment 1 See also Figure 1 , which shows a bearing intelligent diagnosis method in a first embodiment of the present invention, and the method specifically includes steps S101 to S105: S101, collecting a vibration signal, a temperature signal and an acoustic wave signal of a bearing of a target device, and converting the vibration signal, the temperature signal and the acoustic wave signal into corresponding digital signals; For further information, see Figure 2 , the step S101 specifically includes steps S1011~S1012: S1011, performing signal gain adjustment on the vibration signal, the temperature signal, and the sound wave signal to obtain corresponding gain signals; S1012, using a filter to filter each of the gain signals to obtain a corresponding analog signal, and inputting each of the analog signals into a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

[0023] In the specific implementation, several acceleration sensors are installed on the bearing seat cover to generate multiple vibration signals, and a temperature sensor is installed at the bottom of the bearing seat to collect and measure the temperature signal. An acoustic emission sensor is installed on the surface of the bearing seat near the outer ring to generate an acoustic wave signal. Since the bearing is affected by the environment, working conditions, electromagnetic interference, mechanical noise, etc. during operation, resulting in weak vibration signals and high noise, signal conditioning is required before converting the above three signals into digital signals. Signal conditioning includes signal amplification to increase signal strength, filtering some high-frequency noise for the first time through filtering, and then sending the gain-adjusted and filtered signals to the analog-to-digital converter for digitization. In this embodiment, the three analog signal amplifiers use a variable gain amplifier (PGA) to achieve dynamic gain control. The advantage is that it is easy to use and does not require additional peripheral resistors. By applying different high and low levels to the control pin, a variety of different gains can be achieved to adapt to different signal strengths. The model uses an INA849 instrument amplifier, where the gain of the INA849 instrument amplifier is With external resistor The relationship can be expressed as: ; Furthermore, the signal after variable gain adjustment may contain high-frequency noise, and a low-pass filter is required to remove unnecessary high-frequency signals. In this embodiment, the model used by the low-pass filter includes but is not limited to Butterworth filter, Chebyshev filter and Bessel filter, among which the passband phase response of the Bessel filter is approximately linear, that is, the group delay of the filter is basically a constant, which can reduce the nonlinear phase distortion inherent in the IIR filter; the stopband of the Chebyshev filter drops the fastest, and the error between it and the ideal filter is the smallest; the Butterworth filter has the flattest passband, and the stopband drop rate is between the first two filters. In this embodiment, in order to keep the original amplitude of the original signal as much as possible after passing through the filter, a Butterworth low-pass filter is used. Since the Butterworth low-pass filter is the filter damping ratio, then the cut-off frequency The expression is: ; The analog signal after gain and filtering then enters the ADC (digital-to-analog converter) and is converted into a digital signal. The conversion formula is: ; in, is the analog voltage signal output by the sensor. It is the output voltage signal obtained after amplification and filtering. is the output voltage range of the ADC, b is the resolution of the ADC (digital-to-analog converter), and D is the digital signal output.

[0024] S102, inserting new data points at the time points of each of the digital signals to fill in the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals; For further information, see Figure 3 , the step S102 specifically includes steps S1021~S1022: S1021, inserting a new data point at a time point of each of the digital signals to fill in the missing data of each of the digital signals; S1022, using a linear interpolation algorithm to connect known data points in the processed digital signal, and generating estimated values ​​between each of the known data points, so that the processed data signal presents a smooth transition on the time axis, and filtering and denoising the digital signal processed by the linear interpolation algorithm to obtain a corresponding filtered signal.

[0025] In the specific implementation, since the above three signals come from different sensors, their corresponding sampling frequencies and sampling time points are not completely consistent. Therefore, before signal fusion, signal alignment is required. Specifically, missing data is filled in the signal, and new data points can be inserted at the time point of the sparse signal through linear interpolation to align it with other signals; in addition, a smooth alignment method is adopted, and known data points are connected by linear interpolation, and estimated values ​​between these points are generated so that the signal presents a smooth transition on the time axis, which can avoid discontinuous jumps in time alignment and avoid deviations in analysis results due to time misalignment. The calculation formula of the linear interpolation algorithm is: ; In the formula, , represents two known time points in the processed data signal, and ; , Represents the processed data signal at two time points , The value at Indicates the time point where interpolation is required. Represents the signal value calculated by the linear interpolation algorithm.

[0026] Furthermore, the Kalman filter algorithm is used to filter and reduce noise on the above signal. The Kalman filter uses a recursive optimal estimation algorithm to efficiently use the dynamic model and the observation model to recursively estimate the signal after linear interpolation, thereby achieving denoising, smoothing and filtering of the signal. The Kalman filter steps are as follows: 1. Initialize the initial state estimate of the Kalman filter Initial error covariance matrix State transition matrix Control Input Matrix Measurement Matrix Measurement noise covariance matrix Process noise covariance matrix ; 2. The prediction step is to , based on the state estimate at the previous moment, using the system's dynamic model prediction: ; And, update the predicted covariance matrix: ; 3. Update steps Based on the current measurement data , corrected by the Kalman gain: ; ; And update the covariance matrix: ; After Kalman filtering, the random signals in the above signals are effectively removed, making the filtered signals smoother and the main components of the filtered signals more prominent, which is helpful for the subsequent feature extraction.

[0027] S103, extracting features from each of the filter signals, generating a matrix of the extracted feature data, and splicing the obtained data matrix to obtain a multi-source information fusion tensor; For further information, see Figure 4 , the step S103 specifically includes steps S1031~S1032: S1031, performing time-frequency domain feature extraction on each of the filtered signals using a wavelet packet decomposition algorithm to extract time-frequency features of each of the filtered signals in multiple frequency bands; S1032, respectively standardizing the time-frequency features of each of the filtered signals in multiple frequency bands, and performing signal splicing on the standardized signals to obtain a multi-source information fusion tensor.

[0028] In the specific implementation, the wavelet packet decomposition algorithm is used to extract the time-frequency domain features of each filtered signal to extract the time-frequency features of each filtered signal in multiple frequency bands. The advantage of wavelet packet decomposition is that it can efficiently decompose the signal in both the time domain and the frequency domain, thereby realizing more levels of time-frequency analysis, which is helpful for feature extraction, such as mean, standard deviation, kurtosis, skewness and other features, which are used to describe the time-frequency or frequency domain changes of the signal. Among them, the wavelet transform formula is: ; in, Indicates signal Wavelet packet basis functions The signal is decomposed into a low-frequency part and a high-frequency part.

[0029] The decomposition process is as follows: In layer decomposition, the signal It is decomposed into sub-signals of multiple frequency bands, each sub-signal corresponds to a different frequency range. After each layer of decomposition, the signal is further decomposed into low-frequency and high-frequency parts through a filter bank, and the process is repeated until the required decomposition level is reached, thereby obtaining the time-frequency characteristics of the signal in multiple frequency bands.

[0030] Furthermore, the inconsistent dimensions of vibration signals, temperature signals, and acoustic signals will lead to an uneven impact of data on the model when different dimensions are fused, and the inconsistent feature weights will cause high-volatility signals to dominate when fused. Therefore, the three signals are Z-score standardized to reduce the dimension differences for different data. The Z-score standardization steps and formulas are as follows; ; in, It is the original data; is the mean of the data; is the standard deviation of the data; It is the standardized data; The standardization steps are: 1. Calculate the mean: Calculate the mean of each signal (vibration signal, temperature signal and sound wave signal); ; in, is the amount of data; 2. Calculate the standard deviation: Calculate the standard deviation of each signal; ; 3. Standardization: Use the Z-score standardization formula to calculate the standardized data; the standardized data will have the same scale, which is conducive to subsequent model training.

[0031] Specifically, different types of signals are spliced ​​into a large tensor; the method and principle of splicing the large tensor are as follows; The vibration signal, temperature signal and sound wave signal are formed into a high-dimensional data structure, which can simultaneously consider information of multiple dimensions for further processing and analysis. The splicing method is as follows; Among the three signal types mentioned above, the vibration signal shape is ; The shape of the temperature signal is The shape of the sound wave signal is ; After the time series are aligned and have the same time step, the matrix of each signal is generated after filtering, feature extraction and standardization, and then these signals are spliced ​​to obtain a large tensor ; Right now: ; in: Is the size of The vibration signal matrix, Is the size of The temperature signal matrix, Is the size of The concatenated tensor The dimension is .

[0032] S104, using a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, and then inputting the result into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; In the specific implementation, tensor decomposition technology - Tucker decomposition is used to reduce the dimension and fuse the features of multi-source information fusion tensors. The principle and steps of Tucker decomposition are as follows; 1. The specific algorithm of the original tensor decomposition technology-Tucker decomposition is: initialization: , Tucker decomposition decomposes the tensor into a core tensor G and three factor matrices , , .

[0033] ; Among them, G is the core tensor, , , are factor matrices that correspond to the three dimensions of the tensor (number of samples, time steps, and signal type).

[0034] The features after tensor decomposition are divided into data sets, and the data sets are divided into training sets and test sets with a ratio of 80% and 20%. The training set is used as a training model. The data includes three types of faults and bearing features under normal conditions. The three types of faults are inner ring faults, outer ring faults and rolling element faults. Each fault type contains three faults of different sizes. There are 10 different bearing fault features in total. In addition, 4 different loads are set. Select the corresponding data sets for different loads. Select 4 data sets A, B, C, and D from loads 0, 1, 2, and 3 respectively. Each data set contains 10 different bearing fault features. The training set data is input into the input layer, and then passed through the LSTM layer-output layer and the loss function to the optimizer. After the training is completed, the test set is used to evaluate the model to increase the accuracy and robustness of the model.

[0035] S105, performing fault diagnosis on the bearing data to be detected according to the fault diagnosis model to output a corresponding fault diagnosis result.

[0036] In a specific implementation, the bearing data to be detected is obtained, and the bearing data to be detected is input into the above-mentioned fault diagnosis model to output the fault diagnosis result of the bearing data to be detected.

[0037] This embodiment collects vibration signals, temperature signals and acoustic signals for bearings, and fuses them with fault data. It uses multi-source information fusion tensors and LSTM (long short-term memory network) to establish a fault diagnosis model. It uses multi-source information fusion tensors to construct a fused feature tensor from multi-source heterogeneous signals of vibration data signals, temperature data signals and acoustic data signals, and uses LSTM (long short-term memory network) to build and improve the accuracy and robustness of the fault diagnosis model. Combined with historical data query and fault alarm, it improves monitoring accuracy and real-time warning.

[0038] In summary, the bearing intelligent diagnosis method in the above-mentioned embodiment of the present invention uses a multi-source information fusion tensor to construct a fused feature tensor from multi-source heterogeneous signals such as vibration data signals, temperature data signals and acoustic wave data signals. By fusing fault data with real-time data, the accuracy and sensitivity of fault diagnosis can be effectively improved. The obtained multi-source information fusion tensor and the long short-term memory network model are used to construct a fault diagnosis model to improve the accuracy and robustness of the fault diagnosis model. It can not only reflect the current status of the equipment in a timely manner, but also provide real-time alarms, storage and retrieval analysis, thereby identifying potential faults in advance, reducing downtime, optimizing equipment maintenance cycles, and improving equipment operation efficiency and safety. The diagnosis model is continuously optimized through advanced technologies such as machine learning to adapt to the fault diagnosis needs of different types of equipment, and has wide applicability.

[0039] Embodiment 2 Another aspect of the present invention is to provide a bearing intelligent diagnosis system. Figure 5 , shown is a bearing intelligent diagnosis system in a second embodiment of the present invention, the system comprising: The signal acquisition module 11 is used to collect the vibration signal, temperature signal and sound wave signal of the target equipment bearing, and convert the vibration signal, temperature signal and sound wave signal into corresponding digital signals; Furthermore, the signal acquisition module 11 includes: A signal gain unit is used to adjust the signal gain of the vibration signal, the temperature signal and the sound wave signal to obtain corresponding gain signals, wherein the signal gain The corresponding peripheral resistor The relationship is: ; The signal conversion unit is used to filter each of the gain signals using a filter to obtain a corresponding analog signal, and input each of the analog signals to a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

[0040] The signal processing module 12 is used to insert new data points at the time points of each digital signal to fill the missing data of each digital signal, and perform signal alignment and filtering and noise reduction processing on the processed digital signal to obtain a corresponding filtered signal; Furthermore, the signal processing module 12 includes: A signal filling unit, used for inserting a new data point at a time point of each of the digital signals to fill in the missing data of each of the digital signals; The signal processing unit is used to connect the known data points in the processed digital signal using a linear interpolation algorithm, and generate estimated values ​​between each of the known data points so that the processed data signal presents a smooth transition on the time axis, and filter and denoise the digital signal processed by the linear interpolation algorithm to obtain a corresponding filtered signal.

[0041] The feature extraction module 13 is used to extract features from each of the filter signals, generate a matrix of the extracted feature data, and splice the obtained data matrix to obtain a multi-source information fusion tensor; Furthermore, the feature extraction module 13 includes: A feature extraction unit, used to perform time-frequency domain feature extraction on each of the filter signals using a wavelet packet decomposition algorithm, so as to extract the time-frequency features of each of the filter signals in multiple frequency bands; The signal splicing unit is used to standardize the time-frequency characteristics of each of the filtered signals in multiple frequency bands respectively, and to splice the standardized signals to obtain a multi-source information fusion tensor.

[0042] A model training module 14 is used to use a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, and then input it into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; The fault diagnosis module 15 is used to perform fault diagnosis on the bearing data to be detected according to the fault diagnosis model to output corresponding fault diagnosis results.

[0043] The functions or operation steps implemented when the above modules and units are executed are generally the same as those in the above method embodiments, and will not be repeated here.

[0044] The bearing intelligent diagnosis system provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the system embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0045] Embodiment 3 The present invention also provides a computer, see Figure 6 , shown is a computer in the third embodiment of the present invention, including a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, the above-mentioned intelligent bearing diagnosis method is implemented.

[0046] The memory 10 includes at least one type of readable storage medium, which includes a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 10 may be an internal storage unit of a computer, such as a hard disk of the computer. In other embodiments, the memory 10 may also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card, etc. Further, the memory 10 may also include both an internal storage unit of the computer and an external storage device. The memory 10 may be used not only to store application software and various types of data installed in the computer, but also to temporarily store data that has been output or is to be output.

[0047] Among them, in some embodiments, the processor 20 can be an electronic control unit (Electronic Control Unit, abbreviated as ECU, also known as a vehicle computer), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor or other data processing chip, used to run the program code stored in the memory 10 or process data, such as executing access restriction programs, etc.

[0048] It should be pointed out that Figure 6 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0049] The embodiment of the present invention further provides a readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned intelligent bearing diagnosis method is implemented.

[0050] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0051] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0052] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0053] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0054] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A bearing intelligent diagnosis method, characterized in that: include: Collecting vibration signals, temperature signals and acoustic wave signals of the target equipment bearing, and converting the vibration signals, the temperature signals and the acoustic wave signals into corresponding digital signals; Inserting new data points at the time points of each of the digital signals to fill in the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals; Extracting features from each of the filter signals, generating matrices of the extracted feature data, and splicing the obtained data matrices to obtain a multi-source information fusion tensor; Using a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, the multi-source information fusion tensor is input into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; Fault diagnosis is performed on the bearing data to be detected according to the fault diagnosis model to output corresponding fault diagnosis results.

2. The bearing intelligent diagnosis method according to claim 1, characterized in that: The steps of collecting the vibration signal, temperature signal and acoustic wave signal of the target equipment bearing and converting the vibration signal, the temperature signal and the acoustic wave signal into corresponding digital signals include: The vibration signal, the temperature signal and the sound wave signal are subjected to signal gain adjustment to obtain corresponding gain signals, wherein the signal gain The corresponding peripheral resistor The relationship is: ； Each of the gain signals is filtered by a filter to obtain a corresponding analog signal, and each of the analog signals is input into a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

3. The bearing intelligent diagnosis method according to claim 1, characterized in that: The steps of inserting new data points at the time points of each digital signal to fill the missing data of each digital signal, and performing signal alignment and filtering and noise reduction processing on the processed digital signal to obtain a corresponding filtered signal include: Inserting new data points at the time points of each of the digital signals to fill in the missing data of each of the digital signals; A linear interpolation algorithm is used to connect known data points in the processed digital signal, and estimated values ​​are generated between each of the known data points so that the processed data signal presents a smooth transition on the time axis. The digital signal processed by the linear interpolation algorithm is filtered and denoised to obtain a corresponding filtered signal.

4. The bearing intelligent diagnosis method according to claim 3 is characterized in that: The calculation formula of the linear interpolation algorithm is: ； In the formula, , represents two known time points in the processed data signal, and ; , Represents the processed data signal at two time points , The value at Indicates the time point where interpolation is required. Represents the signal value calculated by the linear interpolation algorithm.

5. The bearing intelligent diagnosis method according to claim 1, characterized in that: The steps of extracting features from each of the filter signals, generating a matrix of the extracted feature data, and splicing the obtained data matrix to obtain a multi-source information fusion tensor include: Using a wavelet packet decomposition algorithm to extract time-frequency domain features of each of the filtered signals, so as to extract the time-frequency features of each of the filtered signals in multiple frequency bands; The time-frequency characteristics of each of the filtered signals in multiple frequency bands are standardized respectively, and the standardized signals are spliced ​​to obtain a multi-source information fusion tensor.

6. The bearing intelligent diagnosis method according to claim 5, characterized in that: The calculation formula for the standardization process is: ； ； ； In the formula, Represents the time-frequency characteristics of each filtered signal in multiple frequency bands, Representing time-frequency features The mean of Representing time-frequency features The standard deviation of Representing time-frequency features The number of 7. A bearing intelligent diagnosis system, characterized in that: include: A signal acquisition module, used to collect vibration signals, temperature signals and acoustic wave signals of the target equipment bearing, and convert the vibration signals, temperature signals and acoustic wave signals into corresponding digital signals; A signal processing module, used for inserting new data points at the time points of each of the digital signals to fill the missing data of each of the digital signals, and performing signal alignment and filtering and noise reduction processing on the processed digital signals to obtain corresponding filtered signals; A feature extraction module is used to extract features from each of the filter signals, generate a matrix of the extracted feature data, and splice the obtained data matrix to obtain a multi-source information fusion tensor; A model training module, used to use a tensor decomposition algorithm to reduce the dimension and fuse the features of the multi-source information fusion tensor, and then input it into a pre-built long short-term memory network model for training to obtain a fault diagnosis model; The fault diagnosis module is used to perform fault diagnosis on the bearing data to be detected according to the fault diagnosis model to output corresponding fault diagnosis results.

8. The bearing intelligent diagnosis system according to claim 7, characterized in that: The signal acquisition module comprises: A signal gain unit is used to adjust the signal gain of the vibration signal, the temperature signal and the sound wave signal to obtain corresponding gain signals, wherein the signal gain The corresponding peripheral resistor The relationship is: ； The signal conversion unit is used to filter each of the gain signals using a filter to obtain a corresponding analog signal, and input each of the analog signals to a digital-to-analog converter for signal conversion to obtain a corresponding digital signal.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the bearing intelligent diagnosis method as described in any one of claims 1 to 6 is implemented.

10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the bearing intelligent diagnosis method according to any one of claims 1 to 6 is implemented.

Images